Ryan O'Donnell
Rocco A. Servedio
ABSTRACT
We give new upper and lower bounds on the degree of real multivariate polynomials which sign-represent Boolean functions. Our upper bounds for Boolean formulas yield the first known subexponential time learning algorithms for formulas of superconstant depth. Our lower bounds for constant-depth circuits and intersections of halfspaces are the first new degree lower bounds since 1968, improving results of Minsky and Papert. The lower bounds are proved constructively; we give explicit dual solutions to the necessary linear programs.
- D. Angluin. Queries and concept learning. Machine Learning, 2:319--342, 1988. Google Scholar
Digital Library
- J. Aspnes, R. Beigel, M. Furst, and S. Rudich. The expressive power of voting polynomials. Combinatorica, 14(2):1--14, 1994.Google ScholarCross Ref
- H. Buhrman, R. Cleve, and A. Wigderson. Quantum vs. classical communication and computation. In Proceedings of the 30th Annual ACM Symposium on Theory of Computing, pages 63--68, 1998. Google ScholarDigital Library
- R. Beigel. The polynomial method in circuit complexity. In Proceedings of the Eigth Conference on Structure in Complexity Theory, pages 82--95, 1993.Google ScholarCross Ref
- R. Beigel. Perceptrons, PP, and the Polynomial Hierarchy. Computational Complexity, 4:339--349, 1994. Google ScholarDigital Library
- R. Beigel, N. Reingold, and D. Spielman. PP is closed under intersection. Journal of Computer and System Sciences, 50(2):191--202, 1995. Google ScholarDigital Library
- J. Bruck. Harmonic analysis of polynomial threshold functions. SIAM Journal on Discrete Mathematics, 3(2):168--177, 1990.Google ScholarCross Ref
- E. Cheney. Introduction to approximation theory. McGraw-Hill, New York, New York, 1966.Google Scholar
- Y. Freund. Boosting a weak learning algorithm by majority. Information and Computation, 121(2):256--285, 1995. Google ScholarDigital Library
- M. Goldmann. On the power of a threshold gate at the top. Information Processing Letters, 63(6):287--293, 1997. Google ScholarDigital Library
- A. Hajnal, W. Maass, P. Pudlak, M. Szegedy, and G. Turan. Threshold circuits of bounded depth. Journal of Computer and System Sciences, 46:129--154, 1993. Google ScholarCross Ref
- M. Kearns and U. Vazirani. An introduction to computational learning theory. MIT Press, Cambridge, MA, 1994. Google ScholarDigital Library
- A. Klivans, R. O'Donnell, and R. Servedio. Learning intersections and thresholds of halfspaces. In Proceedings of the Forty-Third Annual Symposium on Foundations of Computer Science, 2002. Google ScholarDigital Library
- A. Klivans and R. Servedio. Learning DNF in time 2Õ(n1/3). In Proceedings of the Thirty-Third Annual Symposium on Theory of Computing, pages 258--265, 2001. Google ScholarDigital Library
- M. Krause and P. Pudlak. Computing Boolean functions by polynomials and threshold circuits. Computational Complexity, 7:346--370, 1998. Google ScholarDigital Library
- N. Littlestone. Learning quickly when irrelevant attributes abound: a new linear-threshold algorithm. Machine Learning, 2:285--318, 1988. Google ScholarCross Ref
- N. Linial, Y. Mansour, and N. Nisan Constant depth circuits, Fourier transform and learnability. Journal of the ACM, 40(3):607--620, 1993. Google ScholarDigital Library
- M. Minsky and S. Papert. Perceptrons: an introduction to computational geometry (expanded edition). MIT Press, Cambridge, MA, 1988.Google Scholar
- D. J. Newman. Rational approximation to |x|. Michigan Mathematical Journal, 11:11--14, 1964.Google ScholarCross Ref
- N. Nisan and M. Szegedy. On the degree of boolean functions as real polynomials. In Proceedings of the Twenty-Fourth Annual Symposium on Theory of Computing, pages 462--467, 1992. Google ScholarDigital Library
- R. Paturi and M. Saks. Approximating threshold circuits by rational functions. Information and Computation, 112(2):257--272, 1994. Google ScholarDigital Library
- M. Saks. Slicing the hypercube. Cambridge University Press, 1993.Google ScholarCross Ref
- D. Sieling. Minimization of decision trees is hard to approximate. Technical Report ECCC Report TR02-054, Electronic colloquium on computational complexity, 2002.Google Scholar
- L. Valiant. A theory of the learnable. Communications of the ACM, 27(11):1134--1142, 1984. Google ScholarDigital Library
- N. Vereshchagin. Lower bounds for perceptrons solving some separation problems and oracle separation of AM from PP. In Proceedings of the Third Annual Israel Symposium on Theory of Computing and Systems, 1995.Google ScholarCross Ref
Index Terms
- New degree bounds for polynomial threshold functions
Recommendations
Polynomial threshold functions and Boolean threshold circuits
We initiate a comprehensive study of the complexity of computing Boolean functions by polynomial threshold functions (PTFs) on general Boolean domains (as opposed to domains such as { 0 , 1 } n or { - 1 , 1 } n that enforce multilinearity). A typical ...
Polynomials that Sign Represent Parity and Descartes' Rule of Signs
A real polynomial P ( X 1 , ... , X n ) sign represents f : A n {0, 1} if for every ( a 1 , ... , a n ) A n , the sign of P ( a 1 , ... , a n ) equals $$(-1)^{f(a_{1}, \ldots ,a_{n})}$$ . Such sign representations are well-studied in computer ...
New degree bounds for polynomial threshold functions
A real multivariate polynomial p(x 1, …, x n ) is said to sign-represent a Boolean function f: {0,1}n →{−1,1} if the sign of p(x) equals f(x) for all inputs x∈{0,1}n . We give new upper and lower bounds on the degree of polynomials which sign-represent ...
Comments